PRIMA Review

download report

Transcript PRIMA Review

The PRIMA facility:
Phase-Referenced Imaging and
Micro-arcsecond Astrometry
Plan
• PRIMA Principle
• Scientific objectives
• Physical limitations
– Off-axis angle
– Limiting magnitude
• Requirements
– Group delay measurement accuracy
– Fringe stabilisation
•
•
•
•
Difficulties
PRIMA system & sub-systems
Observation / calibration / operation strategy
Data reduction
PRIMA motivation
• Main limitation of ground interferometers =
atmospheric turbulence =>
– Fast scrambling of the fringes => snapshots
=> short integration time (~ 50 ms in K)
=> low limiting magnitude (VINCI => K~8 on UT)
– Impossibility to measure the absolute position / phase of
the fringes accurately
• Fringe position (introduced OPD) <=> astrometry
• Fringe phase <=> imaging
• Solutions:
– “Adaptive optics for the piston term” => increase the
limiting magnitude
– Find a phase reference (as quasars in radio astronomy)
=> phase-referenced imaging and differential astrometry
u-v plane and reconstructed PSF
• Image intensity: Iim(a) = IFT ( G(u1 -u2) ) (inverse the Fourier
transform)
with u1 -u2 = baseline vector and G = complex visibility
• Good “synthetic aperture reconstruction” if good u-v
coverage
u-v coverage
(UT 8 hours d=-15º)
6
3
u
5
2
4
4
3
1
2
v
0
0
0
1
1
1
2
2
3
3
0
0
0
0 0
0
4 milli arcsec
1
1
4
5
0
1
2
2
1
This is NOT the u-v plane
This IS the u-v plane
Reconstructed
PSF
8 milli arcsec
K-band
Airy disk
UT
Narrow-angle differential astrometry
• Observe two stars
simultaneously
• Slightly different pointing
directions => DOPD to be
introduced in the interferometer,
between the two beams to get
the fringes
a
DOPD = B . sin a
DOPD
T1
B
T2
• Moreover, the differential
astrometric piston introduced by
the atmosphere is several order
of magnitude lower than the full
piston => these perturbation (of
the measured angle) average to
zero rapidly
~ 30 min for 10” separation and
200 m baseline
Phase-referencing + astrometry
Faint Science Object
Bright Guide Star
DS < 60 arcsec
•
Pick up 2 stars in a 2 arcmin field
– bright star for fringe tracking
– faint object / star
•
DOPD = DS.B +  + OPDturb + OPDint
– OPDint measured by laser
metrology
– OPDturb mean tends to 0
– DOPD measured by VINCI /
AMBER / MIDI / FSU
– DS => object position =>
astrometry
–  => object phase => imaging
•
complex method but very powerfull
– many baselines => many nights
•
synthetic aperture imaging @ 2mas
resolution
astrometry @ 10 mas precision
OPD(t)
OPD(t)
B
OPD(t) -OPD(t) = DS B +  + OPDturb + OPDint
•
The scientific objectives
• General
• Stellar environments
– young stars
– evolved stars
– binaries
• AGNs
• Planets =>
– differential astrometry
– gravitational microlensing
PRIMA goals
• 3 Aims:
– faint object observation (by stabilizing the fringes)
• dual-feed / dual-field : 2’ total FoV (2” FoV for each field)
• K= 10?
13 (guide star) - K= 18?
20 (object) on UTs
• K= 10
16 (object) on ATs
8? (guide star) - K= 15?
– phase-referenced imaging
• accurate (1%) measurement of the visibility modulus and phase
• observation on many baselines
• synthetic aperture reconstruction at 2 mas resolution at 2.2 µm
and 10 mas resolution at 10 µm
– micro-arcsecond differential astrometry
• very accurate extraction of the astrometric phase:
10 µas rms
• 2 perpendicular baselines (2D trajectory)
Scientific objectives - imaging
Accretion disks /
debris disks
Structures of 1AU scale
can be observed:
- up to 1kpc at 2.2 µm and
- up to 100 pc at 10 µm
See O. Chesneau’s &
F. Malbet’s talks
Lynne Allen and Javier Alonso
Stellar
~1
magnetosphere
Accretion disk
radius
(AU)
~50 ~100
Planetesimals
Scientific objectives - AGNs
•
•
•
•
Observation of central core elongation, jets, dust
torus...
Currently ~7 objects observable with MIDI (e.g.
NGC 1068), 0-1 with AMBER
With PRIMA: hopefully >~50 with each => better
sample, better spectral coverage
See W. Jaffe’s talk
Jaffe et al. (2003)
Scientific objectives: Sgr A*
• IR imaging of the matter around the black
hole (see J-U. Pott’s poster)
• 10 µas astrometry of the stars in the central
cusp
• See J-U. Pott’s and H. Bartlo’s talks
Distance R0 = 7.62 +/- 0.32 kpc
QuickTime™ and a
YUV420 codec decompressor
are needed to see this picture.
Scientific objectives: GC flares
•
10 µas astrometry of the galactic center flares
– PRIMA can only give partial information on them (1D measurements <=> 1
baseline)
– if PRIMA can reach the appropriate limiting magnitude (UTs needed, also
because of confusion) and accuracy in 30 min (time scale of flare)
– a better instrument for it would be Gravity
courtesy: F. Eisenhauer (MPE)
Scientific objectives: planets
G. Marcy
• Reflex motion of the star due to planet
presence
• Wobble amplitude proportional to:
– planet Mass
– ( star mass )-2/3
– ( planet period )2/3
– 1 / distance to the star
– amplitude does not depend on orbit
inclination
• Complementarity with radial velocities:
– better for large planets at large distances
– not sensitive to sin(i)
– applicable to (almost) all star types
• Need of long-duration survey
programmes to characterise planets far
from the star
• Need to maintain the accuracy on such
long periods !
• See R. Launhardt’s talk
Scientific objectives: micro-lensing
•
•
Difference in amplification on both images =>
– displacement of total photocenter
Example: M = 10 Msun, impact parameter = 1mas,
rE = 3.2 mas
– maximum photocenter displacement = 1.2 mas
– NOT maximum at closest approach
In case of planet around the lens:
Einstein radius =
3.2 mas
lens
source
– secondary photometric peak and
– more complex shape (3 to 5 images) => imaging and
astrometry
•
But has to work on alerts & needs high limiting
magnitude (K~15-16 on secondary object)
1 mas
2
•
2
total
x
y
The physical limitations and
The scientific requirements
• Physical limitations (more in M. Colavita’s talk)
– Atmospheric anisoplanatism
– Sky coverage
• Scientific requirements
– OPD accuracy for imaging / astrometry
– OPD stabilization for fringe tracking
Atmospheric anisoplanatism 1
slope -2/3
Kolmogorov
spectrum
slope -8/3
slope +4/3
Balloon
measurements
at Paranal
slope -2/3
slope -17/3
slope -8/3
Seeing = 0.66”
at 0.5 µm
 = 10 ms
at 0.6 µm
Atmospheric anisoplanatism 2
• Off-axis fringe tracking <=> anisoplanatic differential OPD
 OPDmeasurement  370.B 2 / 3.


Tobs
for narrow angles ( < 180” UT or 40” AT)
and long total observation time Tobs >> ~100s
for Paranal seeing = 0.66” at 0.5µm, 0 = 10 ms at 0.6µm (L. d’Arcio)
Factor = 300 for Mauna Kea (Shao & Colavita, 1992 A&A 262)
– Increases with star separation
– Decreases with telescope aperture (averaging)
– High impact of seeing quality
• Translates into off-axis maximum angles to limit visibility
losses (< 50 to 90%):
– K-band imaging (2 µm)
 
2
V = V0.exp2. . residual_ OPD  
 
  
• Bright fringe guiding star within 10-20”
– N-band imaging (10 µm)
• Bright fringe guiding star within 2’
Anisoplanatism AT
Anisoplanatism UT
Sky coverage
• Sky coverage <=> limiting magnitude
Accuracy requirements
• Phase-referencing measurable: difference of group delay
DOPD = DS.B +  + OPDturb + OPDint
Fringe sensor astrometry imaging
atmosphere Internal metrology
• Astrometric requirement
– For 2 stars separated by 10” - 0.8”seeing - B=200m => Atmosphere
averages to 10µas rms accuracy in 30 min
– <=> 5nm rms measurement accuracy
• Imaging requirement =>
– dynamic range is important (ratio between typical peak power of a
star in the reconstructed image and the reconstruction noise level)
– DR ~ √M .   D where M = number of independent observations
– DR > 100 and M=100 <=> D   < 0.1 <=> 60nm rms in K
• Ability to do off-axis fringe tracking
The problems / difficluties
More in M. Colavita’s talk
•
•
•
•
•
Air refractive index (ground based facility)
Phase reference stars and calibrators
Time evolving targets
Fringe tracking is not easy
Other instrumental problems
Dispersion and H2O seeing
• Transversal & longitudinal dispersion
• Fringe tracking and observation at different 
• Air index of refraction depends on wavelength =>
– phase delay ≠ group delay
– group delay depends on the observation band
– fringe tracking in K does not maintain the fringes stable in
J / H / N bands
• Air index varies as well with air temperature,
pressure & humidity
– overall air index dominated by dry air
– H2O density varies somewhat independently
– H2O effect is very dispersive in IR (between K and N)
• Remedy: spectral resolution
Refractive index of water vapor (©R. Mathar)
H-band
L-band
K-band
N-band
 [THz]
15
6
3
2
1.5

[µm]
Proper phase references
• We want to do imaging =>
– usually the scientific target is faint =>
• Reference star must be bright (K<10 or 13)
• Bright stars are close and big
– need of long baselines
• => High probability that your guide star is:
• resolved => low visibility
• with resolved structures => non-zero phase
• Phase-referencing cannot disentangle between
target phase and reference phase
• Remedies:
– baseline bootstrapping
– characterize your reference star (stellar type, spectrum,
interferometry) as much as possible prior to observation
– find a faint star close to the reference one to calibrate it
Time and evolving targets
• Phase-referencing works with 2 telescopes at a time
=> Measurements of different u-v points are taken
at different epochs
• Changing the baseline takes time (one day but not
done every day)
• If the object evolves, it is a problem
• Remedies:
– relocate more often (but overheads increase)
– if the “evolution” is periodic (Cepheid, planet), plan the
observations at the same ephemeris time
– have more telescopes and switch from one baseline to
another within one night
• No snap-shot image like with phase cl osure but
better limiting magnitude
Fringe tracking problems
• See Monday’s talk
• Injection stability:
Solutions:
 fast tip-tilt sensing close to the instrument
– Use of monomode optical
 optimize injection before starting
fibers as spatial filter =>
wavefront corrugations and  affects limiting magnitude and efficiency
tip-tilt are transformed into
or you accept a not-perfect fringe lock
photometric fluctuations
– Strehl ratio is not stable at
10 ms timescales
– To measure fringes with
enough accuracy for fringe
tracking, one needs ~ 100
photons at any moment
• Telescope vibrations:
– fast and strong sinusoidal
variations of OPD
– difficult to correct with the
normal OPD loop




“Vibration tracking” (predictive control)
“Manhattan 2” (accelerometers)
laser metrology
active / passive damping
Other instrumental problems
• Baseline calibration:
– baseline should be known at better than < 50µm
– experience on ATs:
• calibration at better than 40µm
• stability ca be better than 120µm
– dedicated calibrations are needed
– stability with time and telescope relocation to be verified
• Telescope differential flexures:
– not seen by the internal metrology
– their effect on dOPD must be very limited or modeled
– differential effect of 2nd order (2 telescopes - 2 stars)
• Mirror irregularities & beam footprints
– non-common paths (metrology/star) to be minimized
– bumps on mirrors should be avoided and mapped
PRIMA Facility
• PRIMA general scheme
• Sub-systems
– Star Separators
– Differential Delay Lines
– Fringe Sensor Units
– Calibration source MARCEL
– End-to-end Metrology
– Control Software and Instrument Software
(PACMAN)
PS
PRIMA
Scheme
PS
SES
SES
Baseline, B
Telescope T1
Metrology end
Metrology end
Star Separator
1
Telescope T2
Star Separator
2
OPD
Controller
System
Delay Line 1
(tracking)
Delay Line 2
(fixed)
B, LgB, AB
Differential
Delay Line
(fixed)
Fringe Sensor Unit B
(PS)
Differential
Delay Line
(fixed)
B, LgB, AB
Metrology
System
DL
Data
storage
DL
Differential
Delay Line
(tracking)
Fringe Sensor Unit A
(or MIDI or AMBER)
(SES)
A, LgA, AA
dOPD
Controller
System
Differential
Delay Line
(fixed)
A, LgA, AA
4 sub-sytems
DS
PRIMA System
Instr.
Star Separators (2 AT & 2 UT)
Fringe Sensor Units (2)
PRIMA Metrology (1)
Differential Delay Lines (4)
Star Separators
•
•
•
•
•
•
•
•
•
•
•
•
Star separation: from PSF up to 2’
Each sub-field =
– 1.5” (UT with DDL - AMBER & PACMAN)
– 2” (UT without DDL - MIDI)
– up to 6” (AT)
Independent tip-tilt & pupil actuators on each beam
10Hz actuation frequency (could be pushed to 50 Hz)
Pupil relay to tunnel center (same as UT)
Chopping / counter-chopping for MIDI
Star splitting for calibration step: 40% - 40%
Star swapping for environment drift calibrations
Symmetrical design for easing calibrations
High mechanical & thermal stability
But: many additional reflections (+8 on AT, +4 on UT)
Installed on AT#3 and AT#4. Under commissioning.
Differential Delay Lines
•
•
•
•
•
•
•
•
•
•
To be used with PACMAN and AMBER, not with MIDI
> 200 Hz bandwidth, < 350 µs pure delay
Push the lab pupil to FSU (4m further than now)
Very stringent requirement on pupil lateral motion
Cat’s eye (3 mirrors, 5 reflections)
2 stage actuator (coarse step motor + piezo on M3)
Internal metrology
M3 can be actuated also in tip-tilt (pupil correction ?)
under vacuum
Preliminary Acceptance Europe: beginning of June
PRIMA Metrology
•
•
•
•
•
Super-heterodyne incremental metrology ( =1.3µm)
Propagation in the central obstruction, from the
instrument to the STS (Retro-reflection behind M9)
Output measurement (dOPD and OPD on one of the
stars) written on reflective memory for the OPD/dOPD
controller
Laser frequency stabilization on I2 at d/<10-8 level
Phase detection: accuracy <1nm rms
Pupil tracking: Custom low noise 4-quadrant detectors
(InGaAs): dd<±100 mm Pk
Working on absolute metrology upgrade
15
Opened loop
Closed loop
10
Frequency noise (MHz)
•
•
5
0
Over 30 min:
Requirement:  < 2 MHz
Open loop:  = 7.4 MHz
p-v = 22 MHz
Closed loop:  = 0.145 MHz
p-v = 0.95 MHz
-5
-10
-15
0
5
10
15
20
Time (min)
25
30
35
Fringe Sensor Units
•
•
•
•
•
•
•
•
•
•
•
ABCD with no OPD scanning (based on polarization)
in K band
OPD and group delay accuracy: < 5nm bias
up to 8kHz measurement frequency
single mode fibers after beam combination
no separate photometric channels
spectral dispersion for group delay
fibers up to cryostat to limit background
fast active injection mirrors for injection
integrated with PRIMET
FSUA and FSUB = twins for astrometry
B
achromatic

light
from T2
p2 & s2
 =   
s1 + s2
p2 - s2| = 90°
p1 + p2
s1 + s2
p1 + p2
A
BC
p1 + p2
s1 + s2
PBS
PBS

p1 + p2
compensator
light
from T1
C
 =   
p1 & s1
s1 + s2
D
Ck

 =   

FSU calibration
Fringe tracking (phase)
fFSU = 1kHz
fOPDC = 2kHz
no tip-tilt
© J. Sahlmann,
N. di Lieto
fFSU = 1kHz
fOPDC = 2kHz
tip-tilt after IFG ~36
mas rms (AT)
PRIMA testbed
•
Testbed needed for:
–
–
•
acceptance tests of FSU (almost finished)
extensive system tests FSU + PRIMET + VLTI environment
Includes:
–
–
–
–
MACAO high order residuals
tip-tilt perturbations
vibrations & other OPD perturbations
(D)DL simulators
•
System tests:
–
–
–
–
–
–
–
–
FSU stability
IFG, BTK, VTK tuning
sensitivity (lim. mag.)
detector read-out optimization
# of spectral channels (3 / 5)
fringe tracking reliability
PACMAN & template tests
calibration optimization
PRIMA Control Software
PACMAN /
AMBER /
MIDI
Observation
preparation –
Templates –
Operation
principles
DL CS
IRIS
PRIMA
Control
Software
FSUA
OPD
ARAL
FITS files
Data
Recorder
MET
Interferometer
Supervisor
Software
FSUB
image stabilisation
AT1
STS 1
AT2
STS 2
DDL CS
dOPD
dOPD
differential
OPD
Controller
OPD
OPD
Controller
pupil stabilisation
RMN for (d)OPD
14 control loops working in parallel
Operation, calibration and
data reduction
•
•
•
•
•
Principle: multiple differential measurement
Typical observation
Critical calibrations
Long term trend analysis
Systematic data reduction and observation
preparation
Multiple differences
• PRIMA = quintuple difference
– 2 telescopes, 2 stars, 2 swaps, metrology/star , 2 moments in time
• Very differents scales:
–
–
–
–
–
500m (metrology path) =>
120m OPD =>
~1cm dOPD =>
~100nm fringe stabilization =>
5nm measurement accuracy => 10-11 ratio to propagation length
• PRIMA challenges:
– very complex system (reliability)
– differences to be done cleanly
– 10µas accuracy requires stability & data logging
•
•
•
•
PRIMA can control some things but not the environment
need to measure / calibrate what is not controlled
need to minimize by operation what cannot be calibrated
need of adapted data analysis and reduction software (PAOS = PRIMA
Astrometric Observation & Software) for long term trends
Critical PRIMA calibrations
• Swapping beams (astrometry) =>
– is needed to reject longitudinal differential effects between
both beams and to “zero” the incremental metrology
– no interruption of PRIMA metrology is allowed
• Injected flux and fiber alignment =>
– no photometric channels is a weakness of the FSU
– relative stability of the 4 FSU fibers has to be measured
• FSU / VLTI spectral calibration =>
– fundamental for the group delay bias / stability
• Baseline calibration =>
– to be known with an accuracy better than 50µm
– dedicated observations / calibrations are needed
• Polarization calibration of the VLTI =>
– potential cyclic errors => dedicated observation mode
Examples of long term trends
• Long term trends = effects than cannot be
calibrated in advance nor measured with enough
accuracy
• Telescope repositioning - baseline calibration
– Need to know the differential baseline at ~50µm for
astrometry at 10µas level
• Telescope differential flexures not monitored by the
PRIMA metrology
– Currently: everything above M9
– Very difficult to model at nm levels
• Mirror irregularities & beam footprints
– PRIMA metrology should follow as close as possible the
star path
• Longitudinal dispersion of air in tunnels:
Astrometric PrEparation Software
developed by the DDL-PAOS Consortium
Data Reduction Software
and Analysis Facility
PAOS Consortium
• Pipeline
– Correction of detector effects + data compression
– Gives an approximate DOPD
• “Morning-after” off-line processing
– Correction of daily effects (dispersion) using an “old”
calibration matrix
– Narrow-baseline calibration
– Gives a better DOPD and angle
• Data Analysis Facility (end of 6-month period)
– Fitting of long term trends & better fitting of daily trends
– Computation of an accurate calibration matrix
• Off-line processing (end of 6-month period)
– Idem as morning after but with updated calibration matrix
Conclusions
• PRIMA is aimed at boosting VLTI performances
(limiting magnitude, imaging) + bringing new feature
(astrometry)
• PRIMA is making VLTI more complex but brings
also solutions to current problems
• PRIMA challenges:
– fringe tracking and limiting magnitude
– long term stability
• Scientific objectives are worth the effort
• ESO will provide tools to reduce data and prepare
observations (see summerschool next year)
• => do not be discouraged and enjoy the challenge !